Sign In

awase

4

Updated: Jul 22, 2026

styleukiyoekimonojapanese

Download

1 variant available

fp16 SafeTensor

awase_rin_v1-000004.safetensors

Half precision, best balance • 109.14 MB

Verified:

Type
LoRA
Stats

71

4

2

Reviews
Published

Jul 22, 2026

Base Model

Illustrious

Training
Steps: 4,152
Epochs: 4
Usage Tips
Clip Skip: 1
Strength: 0.6
Hash
AutoV2
FDE705F906
default creator card background decoration
Followers - 67

67

Likes - 212

212

ep4_w0.6_t0_face.png

AWASE – Japanese Kimono Character LoRA | SDXL

A character LoRA for creating Japanese-inspired characters with calm expressions, narrow eyes, pale skin, and detailed kimono patterns and textures.

Suitable for portraits, upper-body shots, full-body compositions, moonlit scenes, shrines, traditional interiors, and surreal environments. The final appearance may vary depending on the checkpoint used.

This LoRA was trained exclusively on copyright-free materials.

Trigger Word

awase

  • LoRA weight: 0.5–0.7 (0.6 recommended)

  • Base model: SDXL 1.0

  • Recommended checkpoint: rinSoftsketch_v40Niji

  • Steps: 24–32

  • CFG scale: 5–7

  • Resolution: 1024×1024 or 832×1216

  • Sampler: Euler a or DPM++ 2M SDE Karras

  • ADetailer: Optional

  • Hires.fix: 1.5× with R-ESRGAN 4x+ Anime6B